A telling gap is at the heart of most universities’ response to generative artificial intelligence. On one side, policy and integrity rules debate whether using ChatGPT constitutes plagiarism. On the other hand, students and colleagues had built their own working relationships with these tools long before guidance arrived. The policy process ran in one direction; practice ran in another.
This is not a failure of institutional desire. It reflects something more fundamental: GenAI cannot be solved by regulation alone. Universities need not a tighter grip on AI use, but a deeper reckoning with how they live with it – in assessments, feedback and day-to-day teaching. That shift, from regulation to integration, is harder and more important. The only way to make it is to learn from each other.
What cross-border collaboration reveals
Over the past two years, we’ve noticed a striking pattern in our conversations with partners across Europe and beyond. Everywhere, regardless of governance model or digital infrastructure, the same tensions arise: what should students be assessed on, when AI can generate passable answers? What does integrity mean when the line between using a tool and doing your own thinking is genuinely blurred? How do we protect student agency when the easiest path is to outsource cognition to a machine?
- Spotlight guide: Boosting AI literacy across your institution
- An insider’s guide to how students use GenAI tools
- Why AI adoption in universities is really a question of trust
These are structural features of the moment, not nationally specific questions. Yet responses remain institution-by-institution: a ban here, a disclosure form there, a redesigned assessment that cannot be scaled. The same fragmentation shows in administration, where ad hoc adoption leads to inconsistent data standards and unclear accountability. This exposes universities, which hold sensitive student data and, unlike corporations, are expected to model ethical accountability, to an uncomfortable level of risk.
There is a deeper asymmetry, too. Disciplines will have varying capacity to adopt AI, and that capacity will also differ between the Global North and South, with paying students accessing far more capable models than free tiers offer. Equity must stay central to the conversation.
The limits of what we already have
The sector is not short of sound advice. Unesco has published AI competency frameworks for students and teachers, and the UK’s Jisc AI maturity toolkit helps institutions assess their readiness across governance, infrastructure and data management.
These are genuine contributions, but they share a limitation. Most operate at a strategic level, offering principles hard to translate into a seminar, or focus narrowly on technical competencies that miss whole disciplines. They tell institutions what to aim for, with little guidance on how students progress from awareness to ethically grounded capability.
Students, in surveys at our university, have been clear about this gap. They are generally interested in using AI and are not confused about its relevance. What they want is guidance: on fairness, on assessment expectations, on building genuine confidence in their own intellectual capabilities and reassurance that their education is developing something the machine cannot replicate. Alongside this, they want to learn how to use AI responsibly.
FUTURES: seven domains of indispensable human capability
It was precisely this gap, practical rather than theoretical, that prompted us to write the FUTURES framework, published this year by the Higher Education Policy Institute. FUTURES proposes seven domains of human capability that higher education should deliberately cultivate, capacities that generative AI cannot exercise with depth: fluency in AI and digital systems; understanding self and well-being; technology ethics and responsibility; social intelligence; resilience and adaptability; emerging technology awareness; and professional engagement, spanning data literacy, ethical reasoning, communication, critical thinking and active citizenship.
Crucially, FUTURES works across disciplines, supporting critical engagement with narrative and agency in the humanities, questions of transparency in philosophy and intentional choices about human creativity in the arts. We position students as reflective, ethically grounded practitioners rather than passive consumers, and the framework complements rather than replaces existing frameworks. Where the Jisc toolkit asks, “Is our institution ready to adopt AI responsibly?”, FUTURES asks, “Are our students and colleagues developing the human capabilities to make that adoption meaningful?”
What this means for assessment and teaching
Assessment is where the tensions are sharpest. Blanket prohibitions are difficult to enforce and send the wrong signal about what learning is for. If the goal is graduates who think critically, reason ethically and make sound decisions in complex situations, assessment should test exactly those capabilities, not the ability to produce text without AI.
This is a reframe, not a capitulation. The question moves from “Can students produce a piece of work?” to “Can students evaluate, critique, and responsibly integrate AI-assisted work into their own practice?” That is more honest about the world graduates are entering, where using powerful AI tools well, and knowing when not to, is genuinely valuable. AI can also speed up personalised feedback, but the relationship between student and tutor cannot be automated. The goal is to free up time for this relationship.
Globally informed, locally grounded
FUTURES is one national contribution to the global conversation, not a blueprint to be exported wholesale. What works at a research-intensive UK institution will require adaptation at a US community college, a West African university or a South-east Asian polytechnic.
But the underlying argument, that AI strategy must centre on human capability rather than merely manage risk, travels well. These questions require sector-level deliberation informed by multiple stakeholders, not the loudest voices in one country. Frameworks such as the UN Global Digital Compact and the Unesco Recommendations on the Ethics of AI already signal the desire for shared accountability that universities should help shape, not simply receive.
The indispensability argument
Beneath every practical debate sits one question: in a world where AI can generate, automate and synthesise at scale, what makes a human’s education indispensable? The answer is not the production of outputs. It is the critical judgement to know when a well-generated answer is wrong, the ethical reasoning to navigate decisions with human consequences, the social intelligence to build trust and lead, and the resilience to adapt when no model predicted the change.
These are not soft skills tacked on to crowded curricula. They are the core purpose of higher education, now made visible by tools that do everything else very well. The invitation to colleagues across institutions is not to adopt a single framework. It’s to imagine AI strategies that are globally informed yet locally grounded, and to treat that collaborative act as the leadership this moment demands.
Doug Specht is head of the Westminster School of Media and Communication and Gunter Saunders is director of digital capability development and AI leadership at the University of Westminster. They are co-authors of Being Indispensable: Capabilities for a Human-AI World, the ‘FUTURES’ Framework (Hepi, 2026).
If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.
comment